CASPER
Computational and Agentic Scientific Practices, Epistemology, and Reasoning
AI is transforming how we conduct science. Based at The Ohio State University, CASPER pursues this transformation across three fronts: agentic systems for astronomical surveys, computational methods for modern inference, and the epistemic implications of AI-assisted discovery.
Rethinking Scientific Practice in the Age of AI
AI represents a disruptive force in how we conduct scientific research. But disruption brings opportunity. CASPER pursues three interconnected directions: deploying LLMs as agents for large-scale astronomical surveys, advancing computational and statistical methods for modern science, and investigating the epistemic implications of AI-assisted discovery.
Based at The Ohio State University, we leverage OSU's deep involvement in major surveys—Roman, DESI, SDSS-V, and ASAS-SN—to ground our work in real scientific applications. And we ask fundamental questions: what does it mean to understand when AI assists discovery?
CASPER brings together astronomy, physics, computer science, and philosophy to address both the practical challenges and foundational questions that AI raises for science.
Three Interconnected Directions
CASPER pursues research that spans from practical AI deployment to foundational questions about scientific knowledge—each direction informing the others.
Agentic Survey Science
Practical AI at Scale
OSU holds one of the largest footprints in next-generation cosmological surveys. With deep involvement in Roman, DESI, SDSS-V, and ASAS-SN, we are uniquely positioned to develop and deploy agentic systems for survey operations—from instrumentation control to transient classification—moving beyond proofs of concept to practical, grounded AI that advances real science at scale.
Computational & Statistical Methods
Advancing Inference at Scale
Modern surveys generate data at unprecedented scales, demanding new computational approaches. We develop generative models for uncertainty quantification, multimodal foundation models for robust feature extraction, knowledge graphs for optimal targeting, and reinforcement learning for instrument control.
Epistemic Implications
Knowledge About Knowledge
As AI transforms scientific practice, it raises fundamental questions: What does it mean to understand a phenomenon when AI assists discovery? How should we evaluate scientific contributions in an era of automation? We investigate these epistemic implications—inspired by our work on agentic systems—bridging philosophy of science with practical AI development.
Key Investigators
Yuan-Sen Ting
Associate Professor of Astronomy
Chris Hirata
Professor of Physics and Astronomy
David Weinberg
Distinguished University Professor
Klaus Honscheid
Professor of Physics
Paul Martini
Professor of Astronomy
Christopher Kochanek
Professor, Ohio Eminent Scholar
Kris Stanek
Professor of Astronomy
John Beacom
Distinguished Professor of Physics and Astronomy
Todd Thompson
Professor of Astronomy
CASPER Fellowship
We seek postdoctoral researchers who want to work at the intersection of AI and astronomical science. CASPER Fellows contribute to developing agentic systems for large-scale surveys and advancing computational methods for modern astronomy.
Survey Science Integration
Direct involvement with Roman, DESI, SDSS-V, and ASAS-SN collaborations
Computational Focus
Develop AI and statistical methods for real astronomical applications
Interdisciplinary Environment
Collaborate across astronomy, computer science, and philosophy
We're Looking For
- 01
Researchers interested in building AI systems for astronomical surveys
- 02
Those who can bridge computational methods with scientific applications
- 03
Independent thinkers who ask substantive questions about AI in science
- 04
Collaborators who can contribute to practical, grounded research
Get in Touch
Interested in collaborating or learning more about CASPER? We welcome inquiries.
Location
Department of Astronomy
The Ohio State University
Columbus, OH 43210
Affiliations
OSU Astronomy
CCAPP
CASPER© 2026 CASPER Initiative, The Ohio State University